探究美国青少年压力¶
数据故事链接——pythonanywhere¶
数据清洗 ——请看下文¶
题记：你会时常感到压力大吗？是谁让你产生了压力？是老师，是父母，是同学而又或是你的男女朋友呢？¶
说明：此次交互式可视化分析主要想分析影响青少年压力的客观因素。¶
思考¶
- 相关关系而不是因果关系。压力我们一般都会想到的是学习压力大，期末项目多，补课等。但是否可以突发奇想，也有一些我们意想不到的相关性呢？
数据来源¶
- 数据
| 序号 | 数据 | 数据来源 |
|---|---|---|
| 1 | 青少年压力和心理健康问题 | data.world |
| 2 | 美国各州收入 | 美国经济分析局 |
| 3 | 青少年摄入营养 | 联合国粮食及农业组织 |
- 相关整理部分截图（具体请看下文“数据清洗”）
故事结论¶
- 在压力频率调查中，有超过8成的青少年不时会有压力，并且超过4成的青少年一直有压力。
- 根据地区压力大小的对比发现，压力可能与工资水平、成绩绩点，贫困率息息相关。
- 在角色施加压力中，情侣施加的压力比父母、老师都大。可见情感问题会给青少年带来较多困扰。
- 角色压力相关关系比较中，父母压力与老师压力性质大致相似。
- 对于大部分喜欢利用食物缓解压力的人，不喜欢运动缓解压力。
In [31]:
import pandas as pd
pressure = pd.read_csv('./mental_health_poll_updated.csv',encoding='gbk').set_index('问题')
In [69]:
pressure.head(5)
Out[69]:
问题个数¶
In [71]:
Q_type = set(pressure.index)
Q_type
Out[71]:
- 问题一：How often are you stressed 您多久受到一次压力？
- 问题二：What are you most likely to do when you're stressed ? 最让您感到压力的是什么？
- 问题三：What resources do you use to help ? 您使用什么资源来帮助您？
- 问题四：What stresses you out the most ?您最有可能面对什么压力吗？
将Question设为索引¶
提取问题一¶
In [73]:
Q1 = pressure.loc['您多久受到一次压力？']
Q1.head()
Out[73]:
In [74]:
list(set(Q1['分类']))
Out[74]:
In [75]:
len([Q1['分类'].values[i] for i in range(len(Q1['分类'].values)) if Q1['分类'].values[i] == '从不'])
Out[75]:
In [76]:
columns = 'num'
In [77]:
index = list(set(Q1['分类']))
In [78]:
never_num = len([Q1['分类'].values[i] for i in range(len(Q1['分类'].values)) if Q1['分类'].values[i] == '从不'])
never_num
Out[78]:
In [79]:
rarely_num = len([Q1['分类'].values[i] for i in range(len(Q1['分类'].values)) if Q1['分类'].values[i] == '很少'])
rarely_num
Out[79]:
In [80]:
all_num = len([Q1['分类'].values[i] for i in range(len(Q1['分类'].values)) if Q1['分类'].values[i] == '一直'])
all_num
Out[80]:
In [81]:
sometimes = len([Q1['分类'].values[i] for i in range(len(Q1['分类'].values)) if Q1['分类'].values[i] == '有时'])
sometimes
Out[81]:
提取问题一画扇形图¶
In [82]:
often_num = pd.DataFrame([never_num,rarely_num,all_num,sometimes])
often_num
Out[82]:
In [83]:
often_num= often_num.rename(columns={0:'num'})
In [84]:
often_num
Out[84]:
In [85]:
pie1 = often_num.num
pie1_list = [num for num in often_num['num']]
labels = [index for index in often_num.index]
In [86]:
pie1_list
Out[86]:
In [87]:
import plotly as py
import plotly.graph_objs as pygo
py.offline.init_notebook_mode()
fig = {
"data": [
{
"values": pie1_list,
"labels": labels,
"domain": {"x": [0, .5]},
"name": 'How often are you stressed',
"hoverinfo":"label+percent+name",
"hole": .5, ## 同心圆宽度
"type": "pie"
},],
"layout": {
"title":"How often are you stressed",
"annotations": [
{ "font": { "size": 20},
"showarrow": False,
"text": "",
"x": 0.20,
"y": 1
},
]
}
}
py.offline.iplot(fig)
封装函数¶
In [89]:
import plotly as py
import plotly.graph_objs as pygo
py.offline.init_notebook_mode()
pressure = pd.read_csv('./mental_health_poll_updated.csv',encoding='gbk')
Q_type = set(pressure['问题'])
pressure = pressure.set_index('问题')
def question(question):
question = question
for i in range(len(Q_type)):
q = pressure.loc[question]
index = list(set(q['分类']))
num = []
for y in range(len(index)):
type1 = len([q['分类'].values[z] for z in range(len(q['分类'].values)) if q['分类'].values[z] == index[y]])
num.append(type1)
df_num = pd.DataFrame(num)
df_num.index = index
df_num= df_num.rename(columns={0:'num'})
pie = df_num.num
pie_list = [num for num in df_num['num']]
labels = [index for index in df_num.index]
fig = {
"data": [
{
"values": pie_list,
"labels": labels,
"domain": {"x": [0, .5]},
"name": question,
"hoverinfo":"label+percent+name",
"hole": .5, ## 同心圆宽度
"type": "pie"
},],
"layout": {
"title":question,
"annotations": [
{ "font": { "size": 20},
"showarrow": False,
"text": "",
"x": 0.20,
"y": 1
},
]
}
}
return py.offline.iplot(fig)
question("您使用什么资源来帮助您减轻压力")
In [32]:
len(pressure)
Out[32]:
清除地区和城市空值¶
In [33]:
pressure1 = pressure[~pressure['洲'].str.contains('(not set)')]
In [34]:
len(pressure1)
Out[34]:
In [35]:
pressure2 = pressure1[~pressure1['城市'].str.contains('(not set)')]
导入教育满意程度数据¶
In [36]:
poll = pd.read_csv('./pollingData.csv')
poll.head()
Out[36]:
修改地区名与压力调查数据相同名称¶
In [37]:
poll = poll.rename(columns = {'State':'洲'})
查找两个数据集不同的地区¶
In [39]:
set(poll['洲'])^set(pressure2.洲)
Out[39]:
将压力数据的Mississippi修改为Mississipi¶
In [40]:
pressure2.loc[pressure2.洲 == 'Mississippi','洲'] = 'Mississipi'
In [41]:
set(poll['洲'])^set(pressure2.洲)
Out[41]:
抽取百分比数据¶
In [42]:
poll = poll.drop(["Yes","No","Total"],axis = 1)
去除教育满意程度没有的地区¶
In [43]:
pressure2 = pressure2[~pressure2['洲'].isin(['District of Columbia'])]
pressure2.head()
Out[43]:
验证¶
In [44]:
set(poll['洲'])^set(pressure2.洲)
Out[44]:
抽取问题"How often are you stressed 您多久受到一次压力？"¶
In [45]:
often = pressure2.loc['您多久受到一次压力？']
often.head()
Out[45]:
In [46]:
combin_all = pd.pivot_table(often,index=['洲','分类'],values=['分类'],aggfunc='count').rename(columns = {'城市':'num'})
In [47]:
combin_all.head()
Out[47]:
导入经纬度数据¶
In [48]:
jinwei = pd.read_excel("./jinweidu.xlsx")
jinwei.head()
Out[48]:
合并压力数据、教育满意程度以及经纬度数据¶
In [54]:
combin = often.merge(jinwei,left_on='洲',right_on='Region').drop('Region',axis=1)
combin = combin.merge(poll)
combin.head()
Out[54]:
以压力频率做数据透视表¶
In [56]:
combin_all = pd.pivot_table(combin,index=['洲','关键词','lat','lon','PercentYes','PercentNo','分类','代码'],values=['分类'],aggfunc='count').rename(columns = {'城市':'num'})
In [57]:
combin_all
Out[57]:
计算每个频率的在地区占的百分比¶
In [58]:
list0 = list(combin_all['num'])
In [60]:
num_count = list((often.groupby(['洲'])['分类'].count()))
In [61]:
len(num_count)
Out[61]:
In [62]:
list1 = list(often.groupby(['洲'])['分类'].count())
In [63]:
list2=[]
for i in range(len(list1)):
list2.append(list1[i])
list2.append(list1[i])
list2.append(list1[i])
list2.append(list1[i])
print(list2)
In [64]:
len(list2)
Out[64]:
In [28]:
all_the_time.head()
Out[28]:
In [31]:
import plotly.express as px
import plotly
import plotly.graph_objects as go
import plotly.offline as py
import plotly.express as px
plotly.offline.init_notebook_mode(connected=True)
token = 'pk.eyJ1Ijoieml4aWFuMDMiLCJhIjoiY2s0OXozNjUwMGF1OTNlbzExczA5YzBibCJ9.hXI3YY-r4hW7dwR_4YI_DA'
px.scatter(all_the_time,x='All the time',y='PercentYes',size='PercentYes',color='All the time',color_continuous_scale=px.colors.colorbrewer.OrRd)
In [30]:
combin_all['percent'] = [a/b for a,b in zip(list0,list2)]
combin_all.head()
Out[30]:
压力频率与成绩相关关系¶
In [36]:
px.scatter(all_the_time,x='PercentYes',y='All the time',marginal_x='histogram',marginal_y='histogram')
压力等级计算¶
压力大小计算：利用压力频率，将“一直”设置为0.6、“有时候”设置为0.2、“很少”设置为0.15、“从不”设置为0.05算权重。将压力等级分为四个等级，下面是每个等级分布的地区。¶
In [93]:
import pandas as pd
pressure_measure = pd.read_csv("./mental.csv",encoding='gbk')
In [94]:
pressure_level.head()
Out[94]:
根据权重分为四个压力等级，为下拉列表筛选地图¶
In [95]:
jinwei = pd.read_excel('./jinweidu.xlsx')
pressure_jinwei = pressure_measure.merge(jinwei,left_on='洲',right_on='Region').drop('Region',axis=1)
压力较小=pressure_jinwei[(pressure_jinwei["压力程度"]>0)&(pressure_jinwei["压力程度"]<=25)]
压力一般=pressure_jinwei[(pressure_jinwei["压力程度"]>25)&(pressure_jinwei["压力程度"]<=50)]
压力较大=pressure_jinwei[(pressure_jinwei["压力程度"]>50)&(pressure_jinwei["压力程度"]<=70)]
压力很大=pressure_jinwei[(pressure_jinwei["压力程度"]>70)&(pressure_jinwei["压力程度"]<=100)]
element3=['压力较小','压力一般','压力较大','压力很大']
In [101]:
from flask import Flask, render_template, request
import pandas as pd
import cufflinks as cf
import plotly as py
import plotly.graph_objs as go
import plotly.express as px
fig3 = go.Figure(data=go.Choropleth(
locations=pressure_jinwei['代码'],##图像绘制需要地区代码
z=pressure_jinwei['压力程度'].astype(float), ##颜色变化因素
locationmode='USA-states', ## 选择地图
colorscale='purpor',
autocolorscale=False, # hover text
marker_line_color='white', # line markers between states
colorbar_title="压力等级", ## 轴名称
hovertext=pressure_jinwei['洲']
))
fig3.update_layout(
title={
'text':'各地区压力等级' ,## 图表标题设置
'y':0.9,
'x':0.5,
'font':{'family':'SimHei', 'size':25},
'xanchor': 'center',
'yanchor': 'top'},
geo = dict(
scope='usa',
projection=go.layout.geo.Projection(type = 'albers usa'),
showlakes=True, # lakes
lakecolor='rgb(255, 255, 255)'),
)
py.offline.iplot(fig3)